Dj`WddlmZddlZddlmZddlmZddlmZddl m Z ddl m Z ddlmZdd lmZmZdd lmZddlZd d lmZd dlmZd dlmZd dlmZmZm Z d dl!m"Z"gdZ#ddZ$GddeeZ%Gdde%Z&Gdde%Z'y)) namedtupleN)LinearRegression) make_pipeline)StandardScaler)svd)api) add_constant)ABCMetaabstractmethod)solve)_aicc)DTYPE)_BaseStationarityTest)cdiff check_endog)C_canova_hansen_sd_test)CHTest decomposeOCSBTestcdd}|dzdk(}fd} t|tr|dkDsJ |t j |f|z }|fvrd}t|j ||jd |z dkr td |dz}t j||d } |s| dd } t|t| |z} ||| | } tjd| jd z|z } | |z| jd z } | d kDr:| tjgz}t j| j|z} t j | | |f}t j"|d j}t j||d|d|z}|}t|jd D]}t j$||f}| d kDr|d| }|r|dd }tjg|z}t'|| jz|z} |||| |}t)dd}||| ||}|S#ttf$r tdwxYw)a@ Decompose the time series into trend, seasonal, and random components. Parameters ---------- x : np.array, shape=(n_samples,) The time series of which the trend, seasonal, and noise/random components will be extracted. type_: str The type of decomposition that will be performed - 'multiplicative' or 'additive'. We would use 'multiplicative' generally when we see an increasing trend. We use 'additive' when the trend is relatively stable over time. m: int The frequency in terms of number of observations. This behaves similarly to R's frequency for a time series (ts). filter_: np.array, optional (default=None) A filter by which the convolution will be performed. Returns ------- decomposed_tuple : namedtuple A named tuple with ``x``, ``trend``, ``seasonal``, and ``random`` components where ``x`` is the input signal, ``trend`` is the overall trend, ``seasonal`` is the seasonal component, and `random` is the noisy component. The input signal ``x`` can be mostly reconstructed by the other three components with a number of points missing equal to ``m``. Notes ----- This function is generally used in conjunction with :func:`pmdarima.utils.visualization.decomposed_plot`, which plots the decomposed components. Also there is an example script in the ``examples`` folder of the repo and the ``Examples`` section of the docs as well. References ---------- .. [1] Example of decompose using both multiplicative and additive types: https://anomaly.io/seasonal-trend-decomposition-in-r/index.html .. [2] R documentation for decompose: https://www.rdocumentation.org/packages/stats/versions/3.6.1/topics/decompose multiplicativeadditiverr c"k(r||z S||z SN)abrtype_s ^C:\Crop_Prediction\Backend\crop-ai-system\venv\Lib\site-packages\pmdarima/arima/seasonality.py_decomposer_helperz%decompose.._decomposer_helper[s N "q5Lq5Lz 'f' should be a positive integerNz)'type_' can only take values '{}' or '{}'rz)time series has no or less than 2 periodsvalid)mode?axis decomposedzx trend seasonal random) isinstanceint ValueErrorAssertionErrornponesformatshapeconvolverangelenmathceilnanarraytolistreshapenanmean concatenatelistr)xr!mfilter_ris_m_oddr#err_msghalf_mtrendsma_xsdetrend num_seasons pad_lengthbufferm_arrseasonaltempirandomr+decomposed_tuplers ` @r"rr$sd&NHA H=!S!a!e+e''1$-!# ^X..=ABB  Q!DEE!VF KK7 1E cr 63u:. /F 6E2G))S5;;q>1Q67K/U[[^3JA~rvvh&((7>>+f45 JJwa 0 1Ezz%a(//1Hxx)HWf,==>H D 5;;q> "4>>8T"234A~Lj[)CR= ffX F %,,.(61 2E 21e 1)rAr.)selfrAs r"__init__z"_SeasonalStationarityTest.__init__s  q5_- - r$cy)zEstimate the seasonal differencing term. Parameters ---------- x : array-like, shape=(n_samples,) The time series vector. Nr)rUr@s r"#estimate_seasonal_differencing_termz=_SeasonalStationarityTest.estimate_seasonal_differencing_termsr$N)__name__ __module__ __qualname____doc__rVr rXrr$r"rSrSs .   r$rS) metaclassc leZdZdZedddddddd d d d Zfd ZedZedZ dZ xZ S)raConduct a CH test for seasonality. The Canova-Hansen test for seasonal differences. Canova and Hansen (1995) proposed a test statistic for the null hypothesis that the seasonal pattern is stable. The test statistic can be formulated in terms of seasonal dummies or seasonal cycles. The former allows us to identify seasons (e.g. months or quarters) that are not stable, while the latter tests the stability of seasonal cycles (e.g. cycles of period 2 and 4 quarters in quarterly data). [1] Parameters ---------- m : int The seasonal differencing term. For monthly data, e.g., this would be 12. For quarterly, 4, etc. For the Canova-Hansen test to work, ``m`` must exceed 1. Notes ----- This test is generally not used directly, but in conjunction with :func:`pmdarima.arima.nsdiffs`, which directly estimates the number of seasonal differences. References ---------- .. [1] Testing for seasonal stability using the Canova and Hansen test statistic: http://bit.ly/2wKkrZo .. [2] R source code for CH test: https://github.com/robjhyndman/forecast/blob/master/R/arima.R#L148 gv|d?gE)!XU?g3?g[| ?gwِf?gB?gjWW x?g xOԈ@glf@gKnVS@g @c.tt| |yN)rA)superrrV)rUrA __class__s r"rVzCHTest.__init__s fd$q$)r$c|jd}tjt|dzdz t}ttj||dz dzz}t j ||}ttdtj||}||j|z }|j|zjjtj}|jd} t|| |||\} } t!| d } | j#tj$| j&j(kry|j+d } t-| tj.| jd}d |dzz |j1| jj1| jj1| j1| j3j5zS) Nrr r)dtypegY@g?F) with_mean) compute_uvr)r()r3r0r1r-roundr _seas_dummyrrrfitpredictTastypefloat64rrminfinfordepscumsumr identitydotdiagonalsum)wtssnfrecltruncR1lmch residualsFhatauxNeA AtOmfhatAsvFhatsolveds r"_sd_testzCHTest._sd_tests IIaLwwsAEQ;'s3RXXaAI$#6789   Q ' U +    #b#, $,,r** 44)#&&--bjj9]]1  /vr7D!L 9u - 668bhhrxx(,, , ~~1~%y"++iooa.@"ABa1f  133 3 3 FF!CIcc!fXXZ6 6r$c|jd}|dkDsJdtj|dz}tj|d|zftjz}tj }t d|dzD]^}tjd|z|z|z|z |ddd|zdz f<tjd|z|z|z|z |ddd|dz zf<`|ddd|dz fS)Nrr zFThis function is called internally and should not encounter this issuer) r3r0aranger1r9pir5sincos)r@rArxttfmatrrOs r"rhzCHTest._seas_dummy"s( GGAJ1u 88 8YYq\A ww1q5z"RVV+ UUq!a% ?A$&66!b&1*r/A*=#>DQUaK $&66!b&1*r/A*=#>DAQK  ?AvAvIr$c|j|syt|td}|jd}t |j }|d|zdzkry|j ||}|dkrt ||j|dz kDS|dk(rt |dkDS|d k(rt |d kDS|d k(rt |d kDSt |d |dzzkDS)aEstimate the seasonal differencing term. Parameters ---------- x : array-like, shape=(n_samples,) The time series vector. Returns ------- D : int The seasonal differencing term. The CH test defines a set of critical values:: (0.4617146, 0.7479655, 1.0007818, 1.2375350, 1.4625240, 1.6920200, 1.9043096, 2.1169602, 2.3268562, 2.5406922, 2.7391007) For different values of ``m``, the CH statistic is compared to a different critical value, and returns 1 if the computed statistic is greater than the critical value, or 0 if not. rFrdpreserve_seriesr ggd@4gή$@imgCiq\P@gK7?gV-?) _base_caserrr3r-rAr crit_vals)rUr@rxrAchstats r"rXz*CHTest.estimate_seasonal_differencing_termLs.q!  > GGAJ K q1uqy=q!$ 7vq1u 556 6 7v() ) 7v )* * 8v() )6EQ%Z0011r$) rYrZr[r\rrrV staticmethodrrhrX __classcell__rbs@r"rrsc>)Y Y Y Y(I *@6@6D''R.2r$rceZdZdZddddZd fd ZedZeddZedd Z ed Z d Z d Z xZ S)raPerform an OCSB test of seasonality. Compute the Osborn, Chui, Smith, and Birchenhall (OCSB) test for an input time series to determine whether it needs seasonal differencing. The regression equation may include lags of the dependent variable. When ``lag_method`` = "fixed", the lag order is fixed to ``max_lag``; otherwise, ``max_lag`` is the maximum number of lags considered in a lag selection procedure that minimizes the ``lag_method`` criterion, which can be "aic", "bic" or corrected AIC, "aicc". Critical values for the test are based on simulations, which have been smoothed over to produce critical values for all seasonal periods Parameters ---------- m : int The seasonal differencing term. For monthly data, e.g., this would be 12. For quarterly, 4, etc. For the OCSB test to work, ``m`` must exceed 1. lag_method : str, optional (default="aic") The lag method to use. One of ("fixed", "aic", "bic", "aicc"). The metric for assessing model performance after fitting a linear model. max_lag : int, optional (default=3) The maximum lag order to be considered by ``lag_method``. References ---------- .. [1] Osborn DR, Chui APL, Smith J, and Birchenhall CR (1988) "Seasonality and the order of integration for consumption", Oxford Bulletin of Economics and Statistics 50(4):361-377. .. [2] R's forecast::OCSB test source code: https://bit.ly/2QYQHno c|jSr)aicris r"zOCSBTest. 377r$c|jSr)bicrs r"rzOCSBTest.rr$c0t||jdS)NF)rnobsrs r"rzOCSBTest.sE#sxx7r$)rraicccJtt| |||_||_yr`)rarrV lag_methodmax_lag)rUrArrrbs r"rVzOCSBTest.__init__s# h&&+$ r$ctj|}dtjd|dz zd|dz dzzzzdz S)zCompute the OCSB critical valuegmxҿgj>ҿg'*?g!q᢮rg4,F]ko?)r0logexp)rAlog_ms r"_calc_ocsb_crit_valzOCSBTest._calc_ocsb_crit_valsW q  FF:!23{Y&1,7.. //19: :r$c<|jd}|dk(r|j|dStj||dz z|ftjz}t |D]}|||||z|f<|r)|tj |jd}|S)zPerform the TS laggingrr r))r3r<r0r1r9r5isnanany)ylagomit_narxoutrOs r"_do_lagzOCSBTest._do_lags GGAJ !899Q? "ggqC!G}c*+bff4s AC!a% O  rxx}((a(001C r$c~|dkr"tj|jdStj |||S)z>Create the lagged exogenous array used to fit the linear modelr)r0zerosr3rr)rrrs r" _gen_lagszOCSBTest._gen_lagss9 a<88AGGAJ' '7G44r$ct||}|jddk(r tdt|}tj ||}|dkDr||d}|d|jd}t j |t|jd}||d} tj ||d| jdddf} |jt| } | | z } t|} tj | |}| |d} |d| jdddf}|jt|}| |z }tj|| d|jdjdd|d|jdjddf}t j ||jdS)z7Fit the linear model used to compute the test statisticrzThere are no more samples after a first-order seasonal differencing. See http://alkaline-ml.com/pmdarima/seasonal-differencing-issues.html for a more in-depth explanation and potential work-arounds.r'Nqr)methodr ) rr3r.rrsmOLSr rirjr0hstackr<)r@rArry_first_order_diffrylagmfar_fitz4_yz4_lagz4_predsz4z5_yz5_lagz5_predsz5datas r" _fit_ocsbzOCSBTest._fit_ocsbs"!QZ  # #A &! +:  # $!!!S) R<'( A+1771: <+,000="#$'##$6<^djjm^Q=NO>>,v"67 H_Aw##D#.CDzA)*>>,v"67 H_yy  |  $ $R + |  $ $R +   vva""$"//r$ch|j}|j}|j}d}|dkDr|dk7r |j|}g}g}t d|dzD]?} |j||| |} |j| |j|| Atj|jr t dttj |} | dz }|| } |j||||} | j"dd} | dS#t$r t dwxYw#tjj$r4|jtj|jdY)wxYw#tjj$r||} n t dYwxYw) NrfixedzI'%s' is an invalid method. Must be one of ('aic', 'aicc', 'bic', 'fixed')r zAll lag values up to 'maxlag' produced singular matrices. Consider using a longer series, a different lag term or a different test.zXCould not find a solution. Try a longer series, different lag term, or a different test.r')rArr_ic_method_mapKeyErrorr.r5rappendr0linalg LinAlgErrorr9rallr- nanargmintvalues)rUr@rAmaxlagrcrit_regressionicfuncfitsicvalslag_termri best_index regressiontvalss r"_compute_test_statisticz OCSBTest._compute_test_statistics FF A:&G+ G,,V4 DF!!VaZ0 &&..Ax@CKK$MM&+.  &xx##% ")** R\\&12J!^F#:.O *1ff=J""23'RyW G "FGG Gyy,,&MM"&&)KK%&*yy$$ **, !")** *s0D=E>-F10F1c|j|syt|td}|j|}|j |j }t ||kDS)aEstimate the seasonal differencing term. Parameters ---------- x : array-like, shape=(n_samples,) The time series vector. Returns ------- D : int The seasonal differencing term. For different values of ``m``, the OCSB statistic is compared to an estimated critical value, and returns 1 if the computed statistic is greater than the critical value, or 0 if not. rFr)rrrrrrAr-)rUr@statcrit_vals r"rXz,OCSBTest.estimate_seasonal_differencing_term>sX q!  >++A.++DFF34(?##r$)r)T)rYrZr[r\rrVrrrrrrrXrrs@r"rr}s~"H#"8 N :: 55/0/0b5n$r$rr)( collectionsrr7sklearn.linear_modelrsklearn.pipelinersklearn.preprocessingr scipy.linalgr statsmodelsrrstatsmodels.toolsr abcr r numpy.linalgr numpyr0arimar compat.numpyr stationarityr utils.arrayrrr_arimar__all__rrSrrrr$r"rsv# 1*0!*' /..+ {| 5 *C2 &C2LZ$(Z$r$